AEC (Acoustic Echo Cancellation)

UI Element	Purpose
Input node	Connection point for wiring a near-end microphone input to the AEC block.
Reference Input node	Connection point for wiring the room AEC reference input.
Output node	Connection point for wiring the AEC output signal. This signal will have reduced echo and ambient noise with AGC and PEQ processing applied. This is the signal required to send to the far-end.

UI Element	Purpose
Input meter Input Gain Reference meter Ref Gain	Input meters and Gain controls are provided for microphone and reference inputs. Input Gain (MIC) and Ref Gain controls are used to adjust the Input and Reference so that signal levels are about 0 dBr on the meters in the presence of normal speech and room audio. While these Gain controls may be linked, they should not be adjusted during operation as this will adversely affect AEC operation. If all other settings are correct, these controls may be left set to unity gain.
ERL meter ERLE meter TER meter	There are three AEC meters, ERL, ERLE and TER. The ERL meter provides an indication of how well echo is suppressed based on room acoustics: microphone type, placement/sensitivity and loudspeaker location/loudness. While the AEC block is able to work well with little or no acoustic echo suppression, the more acoustic suppression a designer is able to realize, the better active AEC will work without adding any artifacts. A reading of -8 dB indicates 8 dB of echo suppression. The ERLE meter provides an indication of how much additional echo suppression is being achieved by active acoustic echo cancellation. A reading of -20 dB indicates 20 dB of additional echo suppression. The TER meter is just a convenient way for a system designer to see what the total echo reduction is without having to add ERL and ERLE.
AEC Enable	AEC Enable allows AEC to be enabled or disabled. Checking the box enables AEC operation. This control is useful for A/B comparison of performance with and without active echo cancellation. When AEC is not enabled an acoustic model of the room is still acquired and maintained.
NLP	The NLP control adjusts the level of nonlinear processing performed by the AEC block. Primary echo cancellation is accomplished with linear processing where a room is modeled, echo estimated and subtracted from the signal pick up by the microphone. Nonlinear processing deals more with random, lower-level reflections and nonlinearities that the model does not deal with as well. NLP is adjustable as its operation needs to strike a balance between reducing nonlinear echo components and generating unwanted artifacts and adversely affecting double-talk. While this control is linkable, its value should be set by the system designer and it should not be adjusted during operation. Good performance is typically achieved with a setting of 20% to 50%.
AEC Threshold	The AEC Threshold indicator lights when AEC is active. Threshold is typically set so be off only when a microphone is muted.
Min Threshold	The Min Threshold control allows a system designer to set the minimum signal level picked by the microphone that will enable the AEC algorithm and allow it to model a room and apply active acoustic echo cancellation. When a microphone becomes inactive, this prevents the algorithm from diverging from a good room model acquired while the microphone was active or acquiring and inaccurate model before a microphone is active. While linkable, this control should be set by the system designer as part of system calibration and should not be adjusted during operation.
ANR Enabled	The Halogen AEC block includes Ambient Noise Reduction (ANR). The ANR Enable control allows ANR to be enabled or disabled. ANR is enabled when the box is checked.
Reduction	The ANR Reduction control lets the system designer set the level of ambient noise reduction from 0 dB to 20 dB. Ambient noise includes any relatively constant noise source in the room such as fan noise from a projector. The algorithm also suppresses howling. The control is typically set to adequately reduce steady background noise while still sounding natural. Settings of 6 to 10 dB are typical.

Input meter

Input Gain

Reference meter

Ref Gain

Input meters and Gain controls are provided for microphone and reference inputs. Input Gain (MIC) and Ref Gain controls are used to adjust the Input and Reference so that signal levels are about 0 dBr on the meters in the presence of normal speech and room audio. While these Gain controls may be linked, they should not be adjusted during operation as this will adversely affect AEC operation. If all other settings are correct, these controls may be left set to unity gain.

ERL meter

ERLE meter

TER meter

There are three AEC meters, ERL, ERLE and TER. The ERL meter provides an indication of how well echo is suppressed based on room acoustics: microphone type, placement/sensitivity and loudspeaker location/loudness. While the AEC block is able to work well with little or no acoustic echo suppression, the more acoustic suppression a designer is able to realize, the better active AEC will work without adding any artifacts. A reading of -8 dB indicates 8 dB of echo suppression. The ERLE meter provides an indication of how much additional echo suppression is being achieved by active acoustic echo cancellation. A reading of -20 dB indicates 20 dB of additional echo suppression. The TER meter is just a convenient way for a system designer to see what the total echo reduction is without having to add ERL and ERLE.

AEC Enable

AEC Enable allows AEC to be enabled or disabled. Checking the box enables AEC operation. This control is useful for A/B comparison of performance with and without active echo cancellation. When AEC is not enabled an acoustic model of the room is still acquired and maintained.

NLP

The NLP control adjusts the level of nonlinear processing performed by the AEC block. Primary echo cancellation is accomplished with linear processing where a room is modeled, echo estimated and subtracted from the signal pick up by the microphone. Nonlinear processing deals more with random, lower-level reflections and nonlinearities that the model does not deal with as well. NLP is adjustable as its operation needs to strike a balance between reducing nonlinear echo components and generating unwanted artifacts and adversely affecting double-talk. While this control is linkable, its value should be set by the system designer and it should not be adjusted during operation.

Good performance is typically achieved with a setting of 20% to 50%.

AEC Threshold

The AEC Threshold indicator lights when AEC is active.

Threshold is typically set so be off only when a microphone is muted.

Min Threshold

The Min Threshold control allows a system designer to set the minimum signal level picked by the microphone that will enable the AEC algorithm and allow it to model a room and apply active acoustic echo cancellation. When a microphone becomes inactive, this prevents the algorithm from diverging from a good room model acquired while the microphone was active or acquiring and inaccurate model before a microphone is active. While linkable, this control should be set by the system designer as part of system calibration and should not be adjusted during operation.

ANR Enabled

The Halogen AEC block includes Ambient Noise Reduction (ANR). The ANR Enable control allows ANR to be enabled or disabled. ANR is enabled when the box is checked.

Reduction

The ANR Reduction control lets the system designer set the level of ambient noise reduction from 0 dB to 20 dB. Ambient noise includes any relatively constant noise source in the room such as fan noise from a projector. The algorithm also suppresses howling. The control is typically set to adequately reduce steady background noise while still sounding natural.

Settings of 6 to 10 dB are typical.